Progress and Challenges in SI/PI with AI/ML

James L. Drewniak (Fellow, IEEE) received the B.S., M.S., and Ph.D. degrees in electrical engineering from the University of Illinois, Urbana Champaign, IL, USA, in 1985, 1987, and 1991, respectively. He is currently a Curator’s Professor Emeritus with the Missouri University of Science and Technology (Missouri S&T), Rolla, MO, USA. His research interests include electromagnetic compatibility, signal and power integrity, and electronic packaging.

Artificial Intelligence LLMs and Machine Learning algorithms have had a profound impact on progress in many areas of signal integrity and power integrity. However, significant challenges still remain in providing solutions to everyday SI/PI design problems. High-bandwidth memory, facilitated by the thru silicon via (TSV), and rapidly shrinking silicon feature sizes using innovative lithography technologies have given rise to GPU and memory advances enabling this new age of AI/ML. Many innovations in SI/PI are underway using AI/ML for everything from facilitating the productivity of experienced engineers, to addressing some common and significant design challenges. Examples will be given showing the power of AI for dramatically enhancing the productivity of experienced engineers. Progress in applying ML to solve some of the most difficult and time-consuming problems in SI/PI will also be presented. Finally, thoughts and perspectives on AI/ML directions in SI/PI will be discussed.